PLM vs. SLM

If you work with computer-generated holography or programmable light, you’ve likely used (or at least heard of) a spatial light modulator (SLM). Phase Light Modulation (PLM) is often presented as an alternative to SLMs, but that’s not quite accurate.

PLM is a variety of SLM and each serves different roles for different applications.

How do PLM and SLM relate?

A spatial light modulator is any device that imposes a pixel-by-pixel pattern on a beam of light. There are two types of SLM technologies, differentiated by what they modulate and how:

• Amplitude SLMs control pixel-level brightness. The Digital Micromirror Device behind DLP® is a familiar example: each micromirror tilts to send light into or away from the optics.

• Phase SLMs control the phase of the wavefront (the timing of the light) rather than its brightness. These include Liquid-crystal-on-silicon (LCoS) devices and deformable mirrors - the legacy workhorses of the optics field.

PLM is the newest variety of Phase SLM. Developed by Texas Instruments and built on the same platform as the DMD, PLM offers capabilities that LCoS and conventional SLM can’t match.

What makes PLM different from SLM?

A conventional DMD tilts its mirrors to essentially turn the light beam on or off. With PLM, the same micromirror array is operated in piston mode. Each mirror moves up and down in discrete steps to delay (or modulate) the reflected light.

The fact that PLM is reflective, not transmissive, is a critical distinction. In an LCoS, light passes through a liquid-crystal layer, where PLM reflects it off a moving mirror. That reflective nature allows a single PLM multiplex, or manipulate light across wavelength, polarization and angle of incidence simultaneously, rather than one channel at a time. Several independent channels can share the same device, so it can handle more data with the same area and energy.

PLM’s architecture gives it a set of advantages SLM can’t match on every axis at once:

1. Speed: PLM micromirrors switch at a rate of tens of microseconds, compared to the millisecond response of SLM, making it roughly two orders of magnitude faster.
   
   That speed supports dynamic, real-time applications, such as optical switching for AI traffic in data centers, where the switch has to reconfigure faster than traffic patterns. SLM can be too slow to keep up, but a PLM switch can reroute traffic between bursts.
   
   It also enables more operations per second. Consider a laser writing a pattern on a part moving along a production line. As the target shifts, with LCoS, the system is always aiming slightly behind the object. With PLM, the beam can reposition faster than the target changes, tracking in real time instead of chasing.
   

2. Polarization independence: LCoS modulates light in only one polarization state, the component aligned to the liquid crystal’s axis. Light outside that state isn’t modulated, so systems must polarize the beam first, often discarding up to half of the incoming light before it reaches the device.
   
   PLM is polarization-independent: its micromirrors reflect and phase-shift light regardless of polarization, so none is lost to polarization.
   

3. Higher power and wider thermal range: PLM’s reflective architecture handles industrial laser power without heat-induced drift, and it operates across a wide temperature range without the active thermal management liquid crystal requires.

Together, these move PLM out of the lab and into production applications at scale, whether in a factory for sintering, metal 3D printing, or lithography, or in the field with LiDAR.

Is there a tradeoff in using PLM over SLM?

Where PLM outperforms SLM on speed, control and operating range, SLM does have the advantage in bit depth. SLM offers 8-bit phase resolution against PLM’s 4-bit (visible) or 5-bit (near-IR), offering a finer approximation of the ideal wavefront and higher fidelity from a single hologram. Where fine phase resolution is the limiting factor, as in complex holographic imaging, for example, SLM can be the better instrument.

But PLM’s speed narrows the gap. Because it switches so fast, many sub-holograms can be time-multiplexed and averaged into a cleaner result than a single 4-bit frame would allow.

When is PLM the better option?

The clearest use cases for PLM are where LCoS creates a bottleneck: it’s too slow, too polarization-sensitive or unable to tolerate the power or heat. These include:

• Optical switching for data centers and telecom, where routing speed and quick reconfigurability directly affect throughput and conventional switches rely on mechanically aimed mirrors. In fact, PLM can help improve energy efficiency and reduce cooling demand, supporting more sustainable data center operations.

• Free-space optical communication, including ground-to-satellite and deep-space links, where a laser beam must stay precisely pointed across vast distances because even a small pointing error can break the connection entirely. PLM can steer the beam and adapt it in real-time to compensate for atmospheric turbulence that can interfere, maintaining a link that mechanically aimed optics would lose.

• High-speed beam steering and LiDAR, allowing for scanning in arbitrary patterns, rather than fixed lines or planes, with no moving parts.

• Industrial lithography and laser sintering, where high optical power and heat would degrade a liquid-crystal device.

• Holographic and multi-plane projection, including automotive heads-up displays that place real-time depth-of-field information at the driver’s line of sight.

As TI’s Kristofer Oberascher explains, PLM “opens entirely new doors” in programmable lighting across multiple industries, turning once-theoretical applications into reality at scale.

While it may not replace SLM across the board, PLM extends the platform, making a growing set of high-speed, high-power applications possible.